💾 Archived View for gemini.spam.works › mirrors › textfiles › magazines › CUD › cud0660.txt captured on 2022-06-12 at 10:55:37.
-=-=-=-=-=-=-
From <@vm42.cso.uiuc.edu:owner-cudigest@VMD.CSO.UIUC.EDU> Mon Jul 4 00:02:35 1994
Date: Sun, 3 Jul 1994 22:36:00 CDT
Reply-To: TK0JUT2@MVS.CSO.NIU.EDU
Sender: CU-DIGEST list <CUDIGEST%UIUCVMD.bitnet@vm42.cso.uiuc.edu>
Subject: Cu Digest, #6.60
To: Multiple recipients of list CUDIGEST <CUDIGEST%UIUCVMD.bitnet@vm42.cso.uiuc.edu>
Computer underground Digest Sun June 30, 1994 Volume 6 : Issue 60
ISSN 1004-042X
Editors: Jim Thomas and Gordon Meyer (TK0JUT2@NIU.BITNET)
Archivist: Brendan Kehoe
Retiring Shadow Archivist: Stanton McCandlish
Shadow-Archivists: Dan Carosone / Paul Southworth
Ralph Sims / Jyrki Kuoppala
Ian Dickinson
Coptic Idolator: Ephram Shrewdlieu
CONTENTS, #6.60 (Sun, June 30, 1994)
File 1--Open Letter to Veep Al Gore in re New Computer Standard
File 2--PDC'94 CFP-Artifacts session (revised)
File 3--ACM Releases Crypto Study
Cu-Digest is a weekly electronic journal/newsletter. Subscriptions are
available at no cost electronically.
CuD is available as a Usenet newsgroup: comp.society.cu-digest
Or, to subscribe, send a one-line message: SUB CUDIGEST your name
Send it to LISTSERV@UIUCVMD.BITNET or LISTSERV@VMD.CSO.UIUC.EDU
The editors may be contacted by voice (815-753-0303), fax (815-753-6302)
or U.S. mail at: Jim Thomas, Department of Sociology, NIU, DeKalb, IL
60115, USA.
Issues of CuD can also be found in the Usenet comp.society.cu-digest
news group; on CompuServe in DL0 and DL4 of the IBMBBS SIG, DL1 of
LAWSIG, and DL1 of TELECOM; on GEnie in the PF*NPC RT
libraries and in the VIRUS/SECURITY library; from America Online in
the PC Telecom forum under "computing newsletters;"
On Delphi in the General Discussion database of the Internet SIG;
on RIPCO BBS (312) 528-5020 (and via Ripco on internet);
and on Rune Stone BBS (IIRGWHQ) (203) 832-8441.
CuD is also available via Fidonet File Request from
1:11/70; unlisted nodes and points welcome.
EUROPE: from the ComNet in LUXEMBOURG BBS (++352) 466893;
In ITALY: Bits against the Empire BBS: +39-461-980493
UNITED STATES: etext.archive.umich.edu (141.211.164.18) in /pub/CuD/
ftp.eff.org (192.88.144.4) in /pub/Publications/CuD
aql.gatech.edu (128.61.10.53) in /pub/eff/cud/
world.std.com in /src/wuarchive/doc/EFF/Publications/CuD/
uceng.uc.edu in /pub/wuarchive/doc/EFF/Publications/CuD/
wuarchive.wustl.edu in /doc/EFF/Publications/CuD/
EUROPE: nic.funet.fi in pub/doc/cud/ (Finland)
ftp.warwick.ac.uk in pub/cud/ (United Kingdom)
JAPAN: ftp.glocom.ac.jp /mirror/ftp.eff.org/
COMPUTER UNDERGROUND DIGEST is an open forum dedicated to sharing
information among computerists and to the presentation and debate of
diverse views. CuD material may be reprinted for non-profit as long
as the source is cited. Authors hold a presumptive copyright, and
they should be contacted for reprint permission. It is assumed that
non-personal mail to the moderators may be reprinted unless otherwise
specified. Readers are encouraged to submit reasoned articles
relating to computer culture and communication. Articles are
preferred to short responses. Please avoid quoting previous posts
unless absolutely necessary.
DISCLAIMER: The views represented herein do not necessarily represent
the views of the moderators. Digest contributors assume all
responsibility for ensuring that articles submitted do not
violate copyright protections.
----------------------------------------------------------------------
Date: Thu, 23 Jun 1994 17:12:16 -0500 (CDT)
From: Wade Riddick <riddick@JEEVES.LA.UTEXAS.EDU>
Subject: File 1--Open Letter to Veep Al Gore in re New Computer Standard
An Open Letter To Al Gore,
Vice President of the United States of America
A New Computer Standard: Fixing the Flats on the Information Highway
The U.S. must manage the early adoption of industrywide
standards that render emerging technologies compatible with
each other and speed commercial acceptance. Such standards
make it easier for purchasers to experiment with equipment
embodying new technology and reduce the risk of committing to
a technology that quickly becomes obsolete . . .
In the U.S., technological standards are set with little
regard to such issues. Large companies or government agencies
set de facto standards... Unfortunately, none of these
sources of standards has explicit responsibility for managing
the standards process to best promote a new technology.
- Robert Reich1
One important roadblock often missed by policymakers as they
work to lay the foundations of the information super-highway is the
incompatibility that exists among the operating systems and microchips
that will form the highway's roadbed. When the Clinton Administration
opened the telecommunications industry to competition, its goal was
not to limit consumer choice, but rather to broaden choice by
weakening narrow, monopolistic controls over technology and allowing
small private companies to move technology in many different
directions.
None of this will be possible without a common standard to allow
these diverse innovations to interact. Just as the national economy
needs a common currency and a common language in which to conduct
business, so too does the information superhighway need a standard
through which its components can interact. Since the development of
the U.S. Department of Defense's Advanced Research Projects Agency
Network (ARPANET) in the 1960s, the federal government has done an
admirable job establishing network protocols, which are rules needed
for seamless long-distance data transmission between computers.
Without such standards, today's international computer network, known
as the Internet, would not exist.
The U.S. government, however, has not done a good job of
standardizing the basic commands needed to operate computers-the
languages, compilers, operating systems and other instructions
governing the microprocessor (the central processing unit, or CPU,
that is a computer's "brain"). These forms of programming
instructions are the most valuable types of electronic data because
they tell computers how to handle information. If an application
(program) can be transmitted between two different computers but
cannot run on both machines-the current norm in the industry-the
application's value is limited.
Companies like Apple, IBM, Microsoft, Intel and Novell have
little incentive to create truly open or common standards for
operating systems or microchip instructions because each company in
one way or another competes successfully on the basis of differences
in its products. Proprietary standards (where all rights to the
standard are retained by one firm) are one way these companies can
protect their research and development (R&D) costs from reengineering
by competing firms.2
The Problem
Just as the mercantilist nations of the last century forced
their currency on their colonies and used tariff barriers to
discourage trade with other powers, computer makers in the twentieth
century have set standards governing the internal commerce of their
products to the detriment of the competition.3 In the same way that
19th-century Britain bucked the mercantilist trend, maintained a free
trading regime, and lost ground to "freeloading" traders as a result,
IBM defined an open PC standard and bore the costs of maintaining it
while clone makers got a free ride. With no need for heavy R&D
expenses, these companies could undercut IBM prices by a significant
margin.
In the past, proprietary standards have acted as unfair exchange
standards, making it unnecessarily expensive for consumers to move
their investments in data-and particularly software-around from one
platform (operating system) to another. This deters investment, just
as the asset-trapping nature of a command economy or non-convertible
currency was for many years a substantial deterrent to foreign
investment in Eastern Europe.
Consumers have started demanding more compatibility between
systems, but companies have been slow to react. As _The Economist_
put it, "every firm wants a monopoly-and every firm wants to call
it an open standard."4 Recently, corporations have begun
establishing interfirm alliances to allow their systems to support
"multiple personalities" (multiple operating systems). Future IBM
computers will be able to run Mac software, while Apple's new Power PC
will run Windows and OS/2, thanks to the use of translation and
emulation software.5
John Sculley-the ex-CEO of Apple-points out in _Defying Gravity_
that computer designs can no longer be based just on the engineers'
experience of using the system. No one company has the business
expertise to design an entire system in a world where more diverse
products have to be brought to market faster than ever. That speed
requires higher levels of coordination, cooperation and
standardization between companies. The current proliferation of
cross-licensing agreements falls short of a universal standard. The
incentive to sell incompatible platforms is still there; companies
have just decided to rely on translation software that they make,
called microkernels, instead of full-blown operating systems for their
profits. They have failed to break up the operating system into
individual components that can be built by different companies
according to comparative (instead of historical) advantage.
Someday, as happened with railroads and automobiles, a standard
for interchangeable software parts will emerge, either through
government intervention or the natural evolution of a monopoly out of
the market.6 This monopoly will, however, require government
regulation at some point to prevent abuse, as was necessary with the
railroad and telephone empires.
It is often forgotten why, how, and at what cost the national
railroads were unified. According to John Browning, "like
railroads, new information networks are increasingly being built in
large, monolithic chunks, starting from the long distance links and
working down to the local one."7 Long distance links were the last
part of the national rail system to be built, because it took an
immense effort to integrate incompatible regional networks-
particularly in the South where there were only spur lines.8 In fact,
railroads, highways and even computers9 to a certain extent have been
built up regionally with government stimulus and later coordinated
through national structures. Regional and local monopolies had to be
granted so that proposed standards would be self-enforcing, since
where there is incentive to compete, there is incentive to deviate
from the standard and affect the distribution of market share.
Railroads were easy to standardize because the tracks were
originally built with iron rails that wore out quickly. Tracks had to
be rebuilt often, so it was not difficult-given adequate financial
incentive-to rebuild the gauges to a particular width.10 The advent
of steel, because of its durability, might actually have threatened
this standardization. Fortunately, just as steel was replacing iron
in the 1870s and '80s, local railroad companies came together in
regional alliances to standardize gauges and policies for
transcontinental shipping, ending decades of chaos in the industry.
These alliances greatly reduced costs to the consumer and spurred
investment in new railroad technology.
Some railroad companies concerned with standardization feared
the emergence of a monopoly and tried to preserve their independence
by confederating. They borrowed from the American federalist model of
government to create their own tripartite government with a
legislative assembly, executive branch, and judiciary for settling
disputes. This structure balanced competing regional interests
against one another and produced an efficient, egalitarian, state-of-
the-art continental transportation system.11 Since the governing
convention created by these small cartels did not include all rail
companies, nor address all of the public interest, it collapsed when
Jay Gould and others began forming large conglomerates. New,
antidemocratic giants emerged, which Congress then stepped in to
regulate.
Either through market evolution or government intervention, such
a standardization of CPUs and operating systems is inevitable.
According to _The Economist_, the computer industry is rapidly
becoming "a commodity business"12 with all the accompanying industry-
wide conventions. This is occurring in an industry producing goods
with the highest intellectual property content in history (hardly
characteristic of most commodities).
It is possible for government to move in now, avoid further
costs of incompatibility and establish a forward-looking, flexible
standard that will preclude the development of a monopoly and will
reshape the way value is created in the software industry. In the
process, the hyper-competitive aspects of the computer industry that
have served society so well could be preserved. As the National
Performance Review prescribes, government can set clear goals and act
as a catalyst while allowing private actors to move the ball down the
field.
Because of the peculiar nature of information, such a standard
need not be autocratic, nor would setting one be risky. The Japanese
and European efforts to set High-Definition Television (HDTV)
standards flopped because they locked industry into analog hardware
before superior digital technology was ready. Immature technologies
have never been successfully pushed on society. The software industry
has almost the opposite problem-not so much inventing the wheel or
prematurely setting it in stone as constantly having to reinvent it
(in order to operate applications under different systems).13
A computer's instructions are vastly different than the regular
objects that come to mind when standards are discussed. The
instructions CPUs use are virtual; they are not materially dependent
on any particular piece of hardware. As symbols, they can always grow
and be reinterpreted, unlike manufactured products such as metal pipe,
whose dimensions cannot be changed once cast. Corporate planners,
long resistant to the adoption of a standardizing framework, are
beginning to see the adaptability of computer code as an advantage
upon which a new standard could be based. As the senior technical
editor of *BYTE* put it, "the battle is no longer about whether to
layer object-oriented services and emulation systems . . . on a small
kernel . . . nor whether to build an operating system in this style
but how to do the job right."14 The remaining problem is one of
coordination between corporations in getting these new systems to work
together.
The Solution
The essential features of such a system are easily described.
The system could be called DNA, after its biological counterpart which
binds all organic matter into the same competitive framework. While
object orientation15-the way in which commonly used types of data are
paired with the instructions needed to manipulate that data-makes data
transportable and software highly extensible *within* a platform, DNA
would make that operating system and processor object oriented so that
both data *and* software would be transportable across platforms. In
other words, when a processor receives a standard DNA message telling
it to do something like add two numbers or draw a line, it will have a
library available to translate the instruction into the host language
of that particular processor.
Under this system, it would be up to the CPU's manufacturer to
supply the most basic translation libraries, but other firms could
supply add-ons or extensions for functions too complex for the CPU to
execute. This way, market competition could be used to set standards
for new forms of data, instead of having the government mandate
standards for immature technologies. A company marketing a product
which uses a completely novel form of data-say a device for producing
certain odors16-would have an opportunity to create its own standard
for data by marketing a new extension for the DNA system. A
competitor might also market a similar plug-in, and both companies
could compete to gain supporters for their mini-standard. In the end,
the best solution would likely win out. Companies would not have to
worry about maintaining compatibility with an existing base because no
previous software could produce odors.
The uniform interface of DNA would allow individual firms to use
their expertise to replace inefficient system components easily,
thereby broadening the market for their products. If DNA contained a
standard driver for reading keyboard input, for example, and someone
wanted to market a new voice recognition device that would be
compatible with past software, that company could make a substitute
for the keyboard interface that instead uses the firm's voice
recognition hardware. DNA would increase the marketability of the
voice recognition device, because customers could buy the physical
device without having to upgrade their entire software library.
According to *The Economist*, "today all firms need a niche"17
in the computer market-and universal standards can provide the
necessary framework. DNA would not pick winners, but would instead
make it easier for winners to emerge. Systems would be built
component by component on the basis of efficiency, rather than through
political or alliance considerations.
Much DNA code may have to be interpreted on each platform, but
with a common object code standard each platform would be able to do
this in the most efficient manner. If this standard's basic design is
flawed or technology passes it by (since technology moves faster than
anyone's capacity to plan ahead), certain instructions could be
reserved in advance to switch to a completely new, but as yet
unspecified standard.
In the past, companies have objected to the slight performance
degradation caused by interpretation. The Macintosh has been
successful precisely because of the huge "toolbox"18 of standard
commands it makes available to applications. Because programs "call"
these functions in the system, instead of in the application itself,
Apple has managed to reduce program size and smoothly maintain the
system's evolutionary growth path.
Apple's new PowerPC is the first example of a "multiple
personality" PC capable of running under more than one operating
system. The PowerPC uses a new platform and microprocessor, the 601.
To run the old software, which is written for a 68000 microprocessor,
the PowerPC interprets and translates that code to the 601.
Reinterpreting the old 68000 instructions slows things down, but by
rewriting the toolbox to run on the faster new 601, Apple makes up for
that loss. Users see no performance degradation with old software and
see tremendous gains with new software. Most of Apple's competitors
are planning similar interpretation schemes for their new systems.
Since an open standard requires some sort of monopolistic
power, it is clear that if DNA is implemented, companies will no
longer profit from the creation of monolithic operating systems. The
way value is created in the software and hardware industries would be
radically altered under DNA, as shown in Figure 1, but who wants to
make money reinventing the wheel? Real money is made on the cutting
edges of technology, and this technological advancement should
continue to be driven by the free market.
U.S. policymakers must think seriously now about how to keep
American industries globally competitive for the next fifty years. By
2040, no software power will make money reinventing the wheel. In a
world where microprocessor architectures are proliferating instead of
unifying and where technical progress is speeding up in all areas of
science, a DNA-type standard is needed, if for no other reason than to
coordinate the diffusion of technical expertise. Only by making new
technology generic, so that a user can plug it in and go, will the
learning curve needed to use new technologies efficiently be
conquered.
Technology transfer needs to become more automatic. Many
writers, James Dearing among them, have thought of technology transfer
as a "difference-reduction"19 problem-one of trying to get users and
inventors to share the same knowledge about an invention so that the
person in the field knows how to apply it as well as the inventor. In
fact, really useful technology gets put to uses never dreamed of by
its inventors. The problem is how to insulate the information needed
to use new technology from the knowledge of how it works-which
confuses most consumers.
The historical trend in U.S. technological development is clear;
either government or industry will eventually take steps to stop this
continual rebuilding of operating systems from the ground up. The
real issue to be decided in the telecommunications debate is not over
who owns the virtual asphalt or builds the on-ramps. The question is
who will own the resulting computer standard governing the packaging
of information. Any firm which wins control will have a power not
unlike the government's ability to print money: the firm will control
the currency of day-to-day electronic transactions. This fact is
becoming increasingly apparent and important to policymakers.
According to Admiral Bobby Inman and Daniel Burton, "arcane topics
like technical standards . . . that once were viewed as the
responsibility of obscure bureaucrats will increasingly engage public
officials at the highest levels."20
There is already a consensus in the industry as to what features
computers will incorporate in the next decade. It is also clear that
some sort of standard for object code will emerge as well.
Government, though, has several options for the role it can play in
this process: (1) the Commerce Department, perhaps with some
authorizing legislation, could call industry heads together and order
them to set a common object code standard; (2) Commerce could accept
bids from various companies and groups for such a standard; or (3)
finally, the federal government could itself craft a standard with the
help of qualified but disinterested engineers, and then try to force
it upon the industry through the use of government procurement rules,
control over the flow of research and development money or other
economic levers. The recent victory of Microsoft in its case against
Stac Electronics over protecting its operating system indicates that
some reform of the intellectual property laws may be needed as well.
Given the acrimony in the current debate over the definition of
a much-needed encryption (data security) standard, it is difficult to
identify the most politically feasible path for policymakers to follow
in developing common object code standards. There is enough of a
consensus in the industry and among users now to begin the search for
a solution. A serious effort should also be made to reach a consensus
with other industrialized nations, for computers are globally
interconnected to a degree that no other mass consumer product has
been.
Government can prevent a monopoly if it moves now. The unique
nature of information technology would allow a common standard to
develop without locking the industry into risky, immature technologies
and would accelerate rather than hinder innovation. According to
Nicholas Negroponte, director of MIT's Media Lab, "an open systems
approach is likely to foster the most creative energies for new
services and be a vehicle for the most rapid change and evolution."21
Such an approach would simply provide a stable framework within
which businesses could compete on the basis of their expertise and not
on their historical advantage. This is what America's founding
fathers designed federalism to do from the start: balance competing
sectoral and regional interests against one another to spur
competition and development for the benefit of all.
By Wade Riddick
Author Biography
Wade Riddick is a graduate student and National Science Foundation
Fellow in the Department of Government at the University of Texas. He
received his B.A. in English from Louisiana State University. He can
be reached at RIDDICK@JEEVES.LA.UTEXAS.EDU.
Figure 1
Traditional
Microsoft Windows -> Disk / Screen / Memory / Audio / ... -> User
IBM OS/2 -> Disk / Screen / Memory / Audio / ... -> User
Apple Macintosh -> Disk / Screen / Memory / Audio / ... -> User
Currently users have to pick one complete operating system to run.
__________________________________________________________________
New Systems
- Microsoft Windows
/
Microsoft Windows NT -> kernel -- IBM OS/2 - User
\
- Apple Macintosh
- Microsoft Windows
/
Apple/IBM PowerPC -> kernel -- IBM OS/2 -> User
\
- Apple Macintosh
In systems being introduced this year, users have to pick one
company's kernel and then another company's operating system(s).
___________________________________________________________________
DNA Common Standard
Microsoft Apple IBM
( ( )
) ) (
Disk + Screen + Memory + ..... -> User
Under DNA, no one company will make *the* operating system.
___________________________________________________________________
Notes
1 Robert Reich, "The Quiet Path to Technological Preeminence,"